orthy of consideration for near-term networking strategies.
First of all, switches and adapter cards now offer a wider range of data rates--from over 1 Gbps down to 25.6 Mbps--whi
ch means network administrators can match bandwidth to the specific needs of each network segment. Some corporations are finding that this scalability makes ATM an efficient way to replace central mainframes with distributed client/server clusters or to give select workgroups the bandwidth they need for multimedia applications. To decide if the time is right for you to explore this technology, see "Are You Ready for ATM?".
If you decide that ATM is part of your near-term plans, you'll need to assemble the right hardware and software components for this switched-network architecture. The good news is that you no longer need to be Daddy Warbucks to afford ATM switches and adapter cards. Some of the latter, for example, now sell for prices within the range of expensive Ethernet cards (see "Virtually Well Connected").
But don't conclude that economics has been the only stumbling block to widespread acceptance of ATM. There is still a dearth of applications that can take full advantage of ATM's characteris
tics. One big problem is the continuing lack of a standard API that developers can write to for cross-platform applications. We're still waiting for this programming target. But in the meantime, three alternatives exist to help you tie Windows, NetWare, or Unix clients into an ATM network. "Teach Your Apps to Speak ATM" tells you what tools are available for each environment.
In the end, large-scale ATM implementations will not be here tomorrow. A year or two from now is more likely. But the rehearsals for the big time may finally be coming to an end.